An angular-dependent split-window equation for SST retrieval from off-nadir observations

An angular-dependent split-window equation is proposed for determining the Sea Surface Temperature (SST) at any observation angle, including large viewing angles at the image edges of satellite sensors with wide swaths. The proposed equation takes into account the angular dependences of the atmospheric correction and also the emissivity correction. An explicit dependence on the SSE is considered in an independent term. The inclusion of such a term is not common in the current operational SST algorithms but we consider it appropriate taking into account the non-blackness of the sea surface emission for large angles and also the dependence on wind speed. The equation has been adapted to the Moderate Resolution Imaging Spectrometer (MODIS) on board EOS-Terra and EOS- Aqua, with at surface observation angles up to 65deg, and to the Spinning Enhanced Visible and Infrared Imager (SEVIRI) on board MSG, with even larger observation angles. Angular- dependent coefficients were estimated for the atmospheric and emissivity terms using synthetic data generated from a series of cloud-free, latitude-equally distributed radio sounding profiles. The use of the proposed expression requires as input data: at- sensor brightness temperatures for the split-window bands, the observation angle at each pixel, an estimate of the water vapor content and accurate SSE values for both channels. A comparison of the results of the equation for MODIS with in-situ measurements gathered from different NOAA ships and buoys placed around the world showed an accuracy of about plusmn 0.3 K for any observation angle, including off-nadir viewings, while the MODIS operational algorithm leaded to an error of plusmn 0.7 K at large angles. The proposed technique is mainly recommended to retrieve SST from observations at large viewing angles, since similar values are obtained by the current operational algorithms for nadir viewings.